Artificial Intelligence-Based Approaches for Renal Structure Characterization in Computed Tomography Images

基于人工智能的计算机断层扫描图像中肾脏结构表征方法

• 批准号: 10224190
• 负责人: Timothy Lee Kline
• 金额: $ 11.93万
• 依托单位: MAYO CLINIC ROCHESTER
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-08-01 至 2022-10-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10224190
• 关键词: Abdomen; Affect; Aging; Albuminuria; Anatomy; Area; Arteries; Artificial Intelligence; Autosomal Dominant Polycystic Kidney; Award; Biopsy; Chronic Kidney Failure; Clinic; Clinical Research; Collaborations; Communities; Computer Vision Systems; Data; Data Set; Databases; Detection; Development; Disease; Early Diagnosis; Environment; Fibrosis; Funding; Generations; Goals; Gold; Grant; Health; Hepatic Cyst; Hour; Hypertension; Image; Image Analysis; Imaging technology; Individual; K-Series Research Career Programs; Kidney; Kidney Diseases; Machine Learning; Magnetic Resonance Imaging; Measurement; Measures; Methods; Microscopic; National Institute of Diabetes and Digestive and Kidney Diseases; Nephrons; Organ; Outcome; Pathology; Patient Care; Patient imaging; Patients; Polycystic Kidney Diseases; Radiologic Finding; Renal Blood Flow; Reproducibility; Research; Research Personnel; Research Project Grants; Research Proposals; Resources; Risk; Scanning; Semantics; Services; Stenosis; Structure; Surveys; Techniques; Technology; Time; Transplantation; Tubular formation; Visit; Work; X-Ray Computed Tomography; automated analysis; automated image analysis; automated segmentation; base; clinical decision-making; clinical practice; deep learning; density; early detection biomarkers; graft failure; image processing; imaging biomarker; imaging modality; improved; interest; interstitial; kidney biopsy; learning strategy; living kidney donor; member; microscopic imaging; non-invasive imaging; novel; novel imaging technology; personalized decision; precision medicine; prognostic value; programs; radiological imaging; research clinical testing; tool

ABSTRACT The goal of this R03 Small Grant Program for NIDDK is to provide additional funding for Dr. Kline to expand upon his work on his K award and apply his expertise to new image acquisitions and problems related to renal imaging. Dr. Kline’s work has piqued the interest of many internal and external investigators and has led to recent collaborations with Drs. Rule, Denic, and Kim. Together with Dr. Erickson, this new research team has prepared this R03 proposal which takes advantage of the unique expertise of each team member. The focus of this proposal is to bridge the gap between microscopic observations and those assessable non-invasively by radiological imaging. To do this, we have established a unique dataset of renal CT imaging data and corresponding biopsy measured nephron densities. We have also generated a large database of gold-standard segmentation data of kidneys, cortical regions, and medullary pyramids. Using this existing data, we propose to: (i) develop tools for segmentation of kidneys, segmentation of individual medullary pyramids, and imputing missing parts of the kidneys outside of the imaged field-of-view in the CT image, and (ii) to establish imaging biomarkers of early CKD, and correlate macroscopic imaging findings to underlying microscopic structure. This research will be facilitated by Mayo Clinic’s outstanding clinical and research environment dedicated to improving patient care, as well as the Aging Kidney Anatomy Study (PI: Rule), which led to the generation of this unique and well characterized dataset. Dr. Kline’s background in imaging technologies and image processing makes him particularly well suited to perform this research. In addition to the above aims, near the end of this research project Dr. Kline will submit a highly competitive R01 application expanding upon the findings from this research proposal. This proposal will lead to vast improvements to current analysis workflows, as well as an improved understanding of the prognostic power of renal imaging biomarkers. Obtaining this R03 Award will greatly facilitate Dr. Kline’s transition into a prosperous independent researcher focused on developing novel imaging technologies and image analysis techniques for abdominal organ pathologies.

抽象的 NIDDK的R03小赠款计划的目标是为Kline博士提供额外的资金以扩展 在他获得K奖的工作并将其专业知识应用于与肾脏相关的新图像获取和问题 成像。 Kline博士的工作激起了许多内部和外部调查人员的兴趣，并导致了 与Drs的最新合作。规则，否认和金。这个新研究团队与埃里克森博士一起 准备了这个R03提案，以利用每个团队成员的独特专业知识。重点 该提议是弥合微观观察和可通过不创侵入性观察之间的差距 放射学成像。为此，我们建立了一个独特的肾脏CT成像数据数据集和 相应的活检测量的肾单位密度。我们还生成了一个大的金标准数据库 肾脏，皮质区域和髓质金字塔的分割数据。使用此现有数据，我们建议 至：（i）用于分割肾脏的开发工具，单个髓质金字塔的分割和归纳 在CT图像中，在成像的视野外部缺少肾脏的部分，（ii）建立成像 早期CKD的生物标志物，并将宏观成像发现与潜在的显微镜结构相关。这 研究将由Mayo Clinic的杰出临床和研究环境进行准备 改善患者护理以及衰老的肾脏解剖研究（PI：规则），这导致了生成 这个独特且表征良好的数据集。 Kline博士在成像技术和图像方面的背景 处理使他特别适合进行这项研究。除了上述目的，附近 该研究项目的结尾Kline博士将提交竞争激烈的R01申请，以扩展 这项研究建议的结果。该建议将导致当前分析的大大改善 工作流程以及对肾脏成像生物标志物的预后能力的提高理解。 获得此R03奖将极大地支持Kline博士向繁荣的独立研究员的过渡 专注于开发腹部器官的新型成像技术和图像分析技术 病理。